| > A customer having problem with 8GB-raiddisk, is this supported?
Technically, no. But it probably ought to work anyway. Could you
please be more specific about the nature of the problem? Or supply a
/debug=180 trace?
The V2 NFS protocol can not report a file system size of more than 4
Gb. Most clients request this information as a normal part of the
mounting sequence, but I doubt that many clients pay any attention to
it. If the size is more than this, I don't know if the server just
drops the overflow bits or stumbles into some more obnoxious behavior.
A larger size than this would be technically unsupported, though it
probably ought to work if both the server and client are willing and
able to tolerate it. There may also be the possibility of a signed vs.
unsigned bug in either the server or client, which could make the size
limit 2 Gb instead of 4 Gb.
The next potential stumbling point would be if any individual files are
bigger than 4 Gb (or 2 Gb). This is also a limitation of the V2
protocol. If you have any files this big you will simply be unable to
read them using V2 NFS.
Both of these limits are eliminated in the V3 protocol, but UCX does
not yet support that. As far as I know, neither does any other VMS
server implementation. Nor do most clients, except for Digital UNIX
and Sun.
Julie
|
|
Hi, thanks for your reply. I have asked the customer to try this
trace-thing, but meanwhile I have one consoleoutput from the time
where the problem occurred. I would be great if you could help me
to explain the output.
The situation was the following:
The nfs-clients copies data to the 8GB-raiddisk which is
nfs-mounted. Suddenly they get these errors on the consolelogg
and also nfs stops working.
Here is a bit of the console-printout:
0:37:29 -UCX-I-CFS_RDCB, Resource descriptor address 00718000, type 02
10:37:29 %UCX-E-CFS_ULCKNLCKD, Attempt to release resource not previous
10:37:29 -UCX-I-CFS_LOCKSTS, Lock mode 03, 1/178 granted, status 00000988
10:37:29 %UCX-E-CFS_ERROR, Error: CHME Thread is still holding resource
10:37:29 -UCX-I-CFS_THREAD, Service BFS$GETATTR_FH, sequence 1124
10:37:29 -UCX-I-CFS_RDCB, Resource descriptor address 00718000, type 02
10:37:29 %UCX-E-CFS_ULCKNLCKD, Attempt to release resource not previous
10:37:29 -UCX-I-CFS_LOCKSTS, Lock mode 03, 1/177 granted, status 00000988
10:37:29 %UCX-E-CFS_ERROR, Error: CHME Thread is still holding resource
10:37:29 -UCX-I-CFS_THREAD, Service BFS$GETATTR_FH, sequence 1124
10:37:30 -UCX-I-CFS_RDCB, Resource descriptor address 00718000, type 02
10:37:30 %UCX-E-CFS_ULCKNLCKD, Attempt to release resource not previous
10:37:30 -UCX-I-CFS_LOCKSTS, Lock mode 03, 1/176 granted, status 00000988
10:37:30 %UCX-E-CFS_ERROR, Error: CHME Thread is still holding resource
10:37:30 -UCX-I-CFS_THREAD, Service BFS$GETATTR_FH, sequence 1124
10:37:30 -UCX-I-CFS_RDCB, Resource descriptor address 00718000, type 02
10:37:30 %UCX-E-CFS_ULCKNLCKD, Attempt to release resource not previous
10:37:30 -UCX-I-CFS_LOCKSTS, Lock mode 03, 1/175 granted, status 00000988
10:37:30 %UCX-E-CFS_ERROR, Error: CHME Thread is still holding resource
10:37:30 -UCX-I-CFS_THREAD, Service BFS$GETATTR_FH, sequence 1124
Also, do you have any "hints" about how to tune NFS or is it just
the old way that works, try and try again..
/Dan Lundeborg TSC Sweden
|